Predicting the Law Area and Decisions of French Supreme Court Cases
Octavia-Maria Şulea1, Marcos Zampieri2, Mihaela Vela3, Josef van Genabith3,4 1University of Bucharest, Romania 2University of Wolverhampton, United Kingdom 3Saarland University, Germany 4German Research Center for Artificial Intelligence (DFKI), Germany
Abstract
In this paper, we investigate the application of text classification methods to predict the law area and the decision of cases judged by the French Supreme Court. We also investigate the influence of the time period in which a ruling was made over the textual form of the case description and the extent to which it is necessary to mask the judge’s motivation for a ruling to emulate a real-world test scenario. We report results of 96% f1 score in predicting a case ruling, 90% f1 score in predicting the law area of a case, and 75.9% f1 score in estimating the time span when a ruling has been issued using a linear Support Vector Machine (SVM) classifier trained on lexical features.
中文速览
法国最高法院每年产出海量判决文书，如何让机器自动预测案件的裁决结果、所属法律领域以及判决年代，是一个尚未被充分探索的问题。研究者收集了逾12.6万份法国最高法院（Cour de Cassation）判决文本，先通过关键词屏蔽（masking）去除文本中直接暴露答案的词语以模拟真实应用场景，再用词袋（Bag of Words）特征结合线性支持向量机（SVM）分别训练三个分类器。实验结果表明，预测裁决结果的F1值达96%，预测法律领域的F1值达90%，估计判决年代的F1值也有75.9%，均大幅超过随机基线。这项研究填补了法语司法文本自动分析的空白，同时证明经典机器学习方法在法律文本分类中仍具有很强的竞争力，为司法辅助决策系统的实用化提供了有力支撑。
原文 arXiv:1708.01681;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1708.01681v1